i>NddlZddlmZmZmZddlmZddlmZGddeZ y)N)ListOptionalTuple) Lemmatizer)Tokenc eZdZdZededeeeeefffd Zde deefdZ dedeede efd Z d edeed ed eedeef d Z d edeed ed eedeef dZd edeed ed eedeef dZd edeed ed eedeef dZd edeed ed eedeef dZd edeed e ed eedeef dZd edeed e ed eedeef dZd edeed e ed eedeef dZxZS)SpanishLemmatizerzH Spanish rule-based lemmatizer with morph-based rule selection. modereturnc<|dk(rgd}|gfSt||S)Nrule) lemma_ruleslemma_rules_groups lemma_index lemma_exc)superget_lookups_config)clsr required __class__s i/var/www/vps2.regionflexible.com/Desarrollo/venv/lib/python3.12/site-packages/spacy/lang/es/lemmatizer.pyrz$SpanishLemmatizer.get_lookups_config s) 6>XHb> !7-d3 3tokencd|j|jt|jf}||jvr|j|S|j }|j j}t|j}|dvr|jgS|dvr%|jr|dk7r|jgS|gS|j}|jjdj|ij|}| t|}n|dk(rd}n|}|j|t|} |jjdj|g} t|d|z||| | }tt j#|}||j|<|S) N)eolspace) adpcconjintjpartpropnpunctsconjsymxr#rauxverbr lemmatize_)orthposstrmorphcachetextpos_lowerset is_sent_startlookups get_tablegetlist select_rulegetattrdictfromkeys) selfr cache_keystringr,featuresexclemmasrule_posrindexs rrule_lemmatizez SpanishLemmatizer.rule_lemmatizesZZC ,<=  "::i( (jj u{{# & &LLN# #   ""sg~ ''xll$$[155c2>BB6J ?#YFe|!##Hd8n=DLL**=9==hKE;WT<(#:;$F$--/0F & 9 rr,r@c|jjd}||vr.||D]&}t|dj|s!|dcSy)Nrr)r5r6r3issubset)r=r,r@groupsgroups rr9zSpanishLemmatizer.select_ruleBsU''(<= &= $uQx=))(3 8O $rwordrrDc g}g}|jjdj|gD]6\}}tj|dz||} | |k7s&|j | 8g} d|vr|D]~} | j ds| j ds&|jjdjdgD]+\}}| j tj||| -|j| |D]} | |vs|j | t|dkDr|St|dkDr|S|gS)a" Lemmatize an adjective. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. r$ Number=Plurnsaccentsr r5r6r7resubappendendswithextendlen r=rKr@rrDpossible_lemmasselected_lemmasoldnewpossible_lemmaadditional_lemmaslemmas r lemmatize_adjzSpanishLemmatizer.lemmatize_adjJ\ ..}=AA$K 7HCVVC#IsD9N%&&~6 7 H $"1 S!**3/>3J3J33O$(LL$:$:=$I$M$M!2%SS*00S.1QRS S 01$ .E~&&u- .  ! #" "  !A %" "6Mrc|jjdjdgD]\}}||k(s |gcS|gS)a Lemmatize an adverb. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. radverbs)r5r6r7)r=rKr@rrDr\r]s r lemmatize_advzSpanishLemmatizer.lemmatize_advzsJ ..}=AA)RP HCs{u   v rczg}g}|jjdjdgD]\}}||k(s |gcS|jjdjdgD]\}}||k(s |gcS|jjdjdgD]0\}}tj|dz||} |j | 2|j |t |dk(r|St |dkDr/|D]} | |vs|j | t |dk\r|S|SgS)a" Lemmatize a determiner. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. rdetdet_and_pron_fixeddet_and_pron_generalrMrGr5r6r7rSrTrUrX r=rKr@rrDrZr[r\r]r^r`s r lemmatize_detzSpanishLemmatizer.lemmatize_dets_ ..}=AA%L HCs{u   ..}=AA "  HCs{u   ..}=AA "B  3HC VVC#IsD9N  " "> 2  3 t$  1 $" "  !A %( 2E>#**51 2?#q(&&&&Irc g}g}|jjdj|gD]6\}}tj|dz||} | |k7s&|j | 8g} d|vr|D]~} | j ds| j ds&|jjdjdgD]+\}}| j tj||| -|j| |D]} | |vs|j | t|dkDr|St|dkDr|S|gS)a Lemmatize a noun. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. rrMrNrOrPrQrrRrYs rlemmatize_nounz SpanishLemmatizer.lemmatize_nounrbrc2|jjdjdgD]\}}||k(s |gcS|jd}t j d|drt j dd|}t j dd|}|gS) a Lemmatize a numeral. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. rnum,z(\.)([0-9]{3})$rz\.r.)r5r6r7splitrSsearchrT)r=rKr@rrDr\r] splitted_words r lemmatize_numzSpanishLemmatizer.lemmatize_nums  ..}=AA%L HCs{u    3 99'q)9 :66%d+DvvdD$'v rcg}g}|jjdjdgD]\}}||k(s |gcS|jjdjdgD]\}}||k(s |gcS|jjdjdgD]6\}}tj|dz||} | |k7s&|j | 8|j |t |dk(r|St |dkDr/|D]} | |vs|j | t |dk\r|S|SgS)a Lemmatize a pronoun. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. rpronrhrirMrGrjrks rlemmatize_pronz SpanishLemmatizer.lemmatize_pronsg ..}=AA&"M HCs{u   ..}=AA "  HCs{u   ..}=AA "B  7HC VVC#IsD9N%&&~6  7 t$  1 $" "  !A %( 2E>#**51 2?#q(&&&&Irc bd|vr|j||||Sg}g}t|xsd}|jjdj |gD]6\}}t j |dz||} | |k7s&|j| 8|D]} | |vs|j| t|dk(r|D]} |jjdj dgD]J\}}|| vs t| D]2\} } | |k(s | d| |z| | dzdz} | |vs"|j| 4L|jjdj d gD]3\}}|| vs | j||d} | |vs#|j| 5g}|D][} |jjdj d gD]+\}}|jt j ||| -]|j|t|dkDr|St|dkDr|S|gS) a Lemmatize a verb. word (str): The word to lemmatize. features (List[str]): The morphological features as a list of Feat=Val pairs. index (List[str]): The POS-specific lookup list. RETURNS (List[str]): The list of lemmas. PronType=PrsrrrMr voc_alt_1NrG voc_alt_2rQ) lemmatize_verb_pronr-r5r6r7rSrTrUrX enumeratereplacerW)r=rKr@rrDrZr[r\r]r^r`ichar voc_alt_lemmar_s rlemmatize_verbz SpanishLemmatizer.lemmatize_verbGs X %++D(D%H H4:2 ..}=AA$K 7HCVVC#IsD9N%&&~6 7 % .E~&&u- .  1 $) B $ 6 6} E I I!JHCe|'0'7JGAt#s{05bq C%A.0P #0E#9$3$:$:=$I J J!% 6 6} E I I!BHCe|(- c3(B (E1+22=A B B&- KN LL22=AEEiQST KS!((S.)IJ K K 01  ! #" "  !A %" "6Mrc d}g}|}tj||}|st|dkretj|j ddzd|}|j dg|z}tj||}|t|dkre|j j djdgD]\} } tj| | |}|j j djd ij|} | | d } n-|jd |}|j||d hz ||d } g} |D]}|j j djd ij|} | | j| d T|jd |}| j|j||||d | d zd j| zgS)Nz^(.*?)([mts]e|l[aeo]s?|n?os)$rrMrrrQrr)rr{rx ) rSrtrXrTrJr5r6r7r9rrUryjoin)r=rKr@rrD pron_pattpronsr)mr\r]rA verb_lemma pron_lemmasrxs rr~z%SpanishLemmatizer.lemmatize_verb_prons4  IIi &mE a66!''!*s*B5DWWQZL5(E )T*AmE a  ..}=AA)RP *HC66#sD)D *ll$$[155fbAEEdK ?QJ##FH5D,,N++   J   XD,,((599&"EII$OC""3q6*''9""4#6#6tXtU#STU#VW  XS 388K#8899r)__name__ __module__ __qualname____doc__ classmethodr-rrrrrErr9rarerlrnrvryrr~ __classcell__)rs@rr r sU4c4eDItCy4H.I44+E+d3i+Zsd3iHSM..#'9.47.@DS . c.`#'947@DS  c,33#'93473@DS 3 c3j..#'9.47.@DS . c.`#'947@DS  c655#'954rs" (("g: g:r